Soil-inhabiting invertebrates are extremely diverse in form, differ greatly in size and numbers, and are often aggregated in horizontal and vertical distribution. Methods available for assessing populations are all based on representative samples. They include field counting, use of attractant traps or pitfall traps, and assessing numbers of invertebrates in soil samples. Invertebrates can be separated from soil either by wet or dry methods based on physical principles, or by dynamic methods which use stimuli to make them leave the samples. Physical methods of separation include hand sorting of soil, soil washing and flotation, elutriation in an upward flow of water, use of a technique which involves passing soil over plates covered with a grease film in which the invertebrates become embedded, and centrifuging soil samples in water or dense solutions so that invertebrates come to the surface. Dynamic methods suitable for most arthropods include the use of dry funnels with, high controlled temperature and moisture gradients, and wet-funnel methods which are used mainly for nematodes and echytraeid worms. Dry funnels which are easy to use are the most popular technique, but have the drawback that they do not extract eggs or resting stages. Wet funnels, which are more laborious, recover these stages, but also dead individuals. Flotation methods are best for sandy soils, whereas dry funnels are more efficient for peat or clay soils. For most soils, dry funnels are most efficient for the extraction of micro- and mesoarthropods, but flotation is usually preferable for macroarthropods.